Interface SplitIterator

  • All Known Implementing Classes:
    CompressionSplitIterator, SingleSplitIterator

    public interface SplitIterator
    A forward-only iterator over data splits with associated locality information. The provided locality information can be used to help decide how to assign splits to cluster nodes for processing.
    See Also:
    DataSplit
    • Method Detail

      • next

        boolean next()
              throws IOException
        Advance to the next data split in the iterated set.
        Returns:
        true if there is another split, otherwise false.
        Throws:
        IOException
      • getSplit

        DataSplit getSplit()
        Get the current split in the iterated set.
        Returns:
        the split currently selected
      • getLocalityInfo

        List<String> getLocalityInfo()
        Get the list of machines for which access to the current split is local. An empty list indicates the split is local to no node (or all nodes).
        Returns:
        the machines where the currently selected split is local. If the split is local to no machine (or to all), this list is empty.